Let AI models trade stocks autonomously with $1k each
Poster ran four AI models on a weekday pre-market prompt to trade $1,000 portfolios each for nine weeks; ChatGPT came out ahead.
Done withGemini
- What they did
- The poster set up four paper-trading portfolios with $1,000 each on Alpaca and ran the same weekday pre-market prompt against ChatGPT, Gemini, Claude, and Perplexity with Deep Research enabled. Each model could only return BUY, SELL, HOLD, or CANCEL, the trading flow was automated in Python and logged publicly, and the poster says they did not override any decisions.
- How it went
- After nine weeks, ChatGPT led at +21.1%, helped by healthcare picks including IOVA doubling and ACHC rising 52%. Perplexity was +1.1%, Gemini -6.6%, Claude -11.5%, and the poster stressed it was paper money and too early to trust blindly.
- Worth knowing
- One practical catch from the thread is that ChatGPT sat entirely in cash for almost three weeks before making trades, so the system may do nothing for long stretches.
Try it yourself with Gemini
Every weekday before the market opens, review a paper-trading portfolio of $[1000] and suggest trades based on [your strategy or risk level], using only simulated money. Record each decision with a one-sentence reason and the portfolio value in a running log. After [number] weeks, summarize performance against [benchmark, e.g. S&P 500]. Do not place any real orders; show me any proposed trade for approval first.
Source: Reddit · Mar 12, 2026
Five of these in your inbox every morning
The best things people got an AI agent to do, each with the prompt to try it.
More like this
Race three models to build a Counter-Strike clone
InstantDB had Codex, Opus, and Gemini each try to build a multiplayer shooter.
Turn a blog post into an animated infographic
A content creator fed a blog post into Gemini Canvas, which produced a coded animated infographic using brand colors and audience details from the chat thread.
Reverse engineer legacy systems without source code
Thoughtworks consultants used Gemini 2.5 Pro in a two-week pilot to produce validated blueprints of legacy systems' functionality.
- Custom agent
Shop for umbrella insurance through your personal agent
Personal AI agents use Coverage Cat's Agent API/MCP to size umbrella coverage and compare carrier options with price ranges, backed by a licensed broker.
Negotiate bills, handle returns and cancel subscriptions
A Meta employee said Muse negotiated an AT&T bill, found Amazon returns, returned IKEA furniture and filed a vet claim; another had a forgotten subscription cancelled and refunded.
Track investments over time via Claude Code and Robinhood MCP
Poster asks how to get Claude Code connected to Robinhood via MCP to keep tracking investments after researching or trading; a question, not a finished result.